So Beesley's mobile red coin HS (638) (https://www.speedrun.com/crossy/run/zxdwlv5y) includes a coin doubling glitch that as far as I can tell from discussion with other players is mobile only and random, and is random ie cannot be purposefully induced. I think that these runs should have their scores halved so the score reflects the actual number of coins collected, rather than the difference between ending and starting amounts. Also seeking verification that this is allowed, and was not an oversight or some such by mods.

All righty..... here we go I was watching this run - amazed as most, but more than a few things bothered me.

1. Maths - CR is 60fps, 25 hops requires 24 frames from the 1st hop. This equates to 0.400 seconds. This is assuming double clicks are impossible, which I believe they are. I have gotten one (see last points of 1000 Pac-man run), but I assume it is lag induced. One frame below this is Hexl's run, which going frame by frame, includes several doubles in a frame, as well as not gaining points on a frame.

2. Scroll wheel better than autoclicker? Using an autoclicker to my knowledge should have the same effect as rerouting left click to the scroll wheel. However, I was unable to obtain double clicks, and gained extremely consistent 0.400s using an autoclicker running at more than 15 clicks per frame. I note this based on Hexl's own reply to 'How?' of "scroll wheel"

Now the stuff other people probably didn't see... I found these after finding nothing contradictory to 1. and 2.

3. Audio file contains ending sound during run. This is also clearly audible on the video itself. I maybe to cover up janky sound from a double click? Maybe a bug?

4. Character turn animation: Character turn animation is separate from forwards movement. If you use an autoclicker and compare the turn animation the fox between your's and Hexl's, you should see a noticeable difference. You should also see a disturbing similarity between one forwards input (one frame press and release), and the turn animation of Hexl's fox - around 11/12 frames before it is pretty much straight, whereas an autoclicker run, the fox is about only halfway through turning in the same number of frames.

I will say in Hexl's defence that he dies on the correct lane, cheating if it did occur was not that blatantly obvious.

Id love any input, and would welcome being disproven. Otherwise, I firmly believe that 0.400s is the absolute maximum for this category, even with click + button tech, which I have tested - any release of both on the same frame causes only one input.

I have not yet looked at the 500 category, but I hope I don't find anything similar there. Edit: 50 seems to be at current theoretical maximum, double 25, + 1 frame as no. 26 is included, but score 1 isnt, I see no need to investigate further

I dunno if this one should really be here. Sure, its funny, but (at least on PC) its a matter of just grinding till you get a good start, then pressing the button twice.

This isn't my main problem with it though. The speed of the run at the highest level is entirely dependant on how fast your computer can process the death. I used an autoclicker to get the fastest time possible on my hardware and it is still .7-.8 slower than WR, due to either the mods timing methods varying, or just the difference in computing power.

DragonTamerBrownie, the WR holder of the PC 25 score speedrun has admitted to using an autoclicker for his run and has not been removed. Does this mean that an autoclicker is allowed, or rather, not considered cheating? I personally find this insulting to the people who actually put in the hours to get a good run in the category. If it is allowed in this category, is it also allowed in other categories, because it would have a major impact on 50 and 100 score speedruns as well.